Table 2 N2/N3 NREM vs. wakefulness classification in PD participants

From: Multi-night cortico-basal recordings reveal mechanisms of NREM slow-wave suppression and spontaneous awakenings in Parkinson’s disease

 

Cortex

Subcortex (STN/GPi)

 

30 s epoch

5 s epoch

30 s epoch

5 s epoch

 

5-fold CV

2-fold CV

5-fold CV

2-fold CV

5-fold CV

2-fold CV

5-fold CV

2-fold CV

Accuracy

93.60 ± 1.40

93.65 ± 1.37

90.60 ± 1.63

90.93 ± 1.61

87.43 ± 3.04

87.15 ± 3.09

80.62 ± 4.08

81.80 ± 3.62

AUC

97.50 ± 0.67

97.53 ± 0.69

95.42 ± 1.10

95.70 ± 1.08

94.02 ± 1.83

93.93 ± 1.85

88.30 ± 3.37

89.23 ± 2.97

Sensitivity

93.88 ± 1.67

93.50 ± 1.14

93.62 ± 1.23

90.70 ± 1.47

86.50 ± 3.26

88.00 ± 2.93

88.72 ± 1.36

82.48 ± 4.66

Specificity

93.30 ± 1.15

93.88 ± 1.60

87.57 ± 2.06

91.13 ± 1.79

88.35 ± 2.85

86.33 ± 3.25

72.53 ± 8.14

81.08 ± 2.86

PPV

93.35 ± 1.19

93.85 ± 1.58

88.32 ± 1.87

91.13 ± 1.74

88.07 ± 2.93

86.55 ± 3.17

77.38 ± 5.13

81.18 ± 3.11

NPV

93.88 ± 1.63

93.48 ± 1.17

93.20 ± 1.38

90.73 ± 1.48

86.80 ± 3.17

87.75 ± 3.02

86.22 ± 1.70

82.50 ± 4.16

  1. Individual machine-learning model performance (mean ± SEM) for N2/N3 NREM vs. wakefulness binary classification using cortical data.
  2. PPV positive predictive value, NPV negative predictive value, AUC area under the receiver operating characteristic curve, CV cross-validation.
Back to article page